DXOmark sensor rating

[Octarine]

Imho, that's the key page: http://www.dxomark.com/index.php/en...-DxOMark-Scores/Sensor-Scores/Use-Case-Scores
3 different scenarios, finally normalized and averaged. When comparing DXO with other reviews (e.g. dpreview) it would be more beneficial using the single test results and comparing them rather than using the final DXO rating.

 

[LBL2009]

Imho, that's the key page: http://www.dxomark.com/index.php/en...-DxOMark-Scores/Sensor-Scores/Use-Case-Scores
3 different scenarios, finally normalized and averaged. When comparing DXO with other reviews (e.g. dpreview) it would be more beneficial using the single test results and comparing them rather than using the final DXO rating.

Yes, the use-case-scores should be used for meaningful comparison.

I shoot mostly landscape. Dynamic range is important to me.

 

[2100]

This thread is great, I still remember imaging-resource way back then like more than 10 years ago. You guys still remember the good old stuff like Canon S10 or S20? I was drooling at the Canon D30 Nikon D100 after that. You whip out a digital camera and people go oohh---ahh... I captured my own proposal to my wife in 2001 in South Korea on Yong Pyong mountain at minus 19 degrees C. (I know coz I got thermometer nd captured the image). It was great, NiMH ran of juice, and I just ran to the convenience store and bought 2 packs of 4XAA (if i remember correct it was about S$3.20 for Energizer). And each set allowed me to take about 80 pix only.

And nowadays you still get guys asking the same question, will my $4000 digital SLR survive 0 degrees? Will dust affect stuff? What will moire do. What will noise do
.

 

[coolthought]

Yes, the use-case-scores should be used for meaningful comparison.

I shoot mostly landscape. Dynamic range is important to me.

How will having a better dynamic range helps? Thanks.

 

[alantkh]

I think it is great. The internet is filled with a LOT of misinformation. Example, people looking at high ISO noise performance with a uniform grey patch patch.... I mean it is totally pointless looking at noise level without signal level when discussing SNR.

Dynamic range is also very hard for the layman to measure accurately and 99% of the time the tester is biased for his camera. Dxomark at least have an objective rating and a well documented process which is very good. Their measurement process is probably 99% more accurate and rigorous than what the average internet poster can do.

 

[night86mare]

How will having a better dynamic range helps? Thanks.

sunset timing, DR needed is wider?

granted, you might still need to use tools like multiple exposures for blending, and gnd.... but it offers opportunities to just do this with a single shot with gnd alone (more often).

 

[coolthought]

sunset timing, DR needed is wider?

granted, you might still need to use tools like multiple exposures for blending, and gnd.... but it offers opportunities to just do this with a single shot with gnd alone (more often).

Now you are making the K-5 looks really tempting....

 

[night86mare]

Now you are making the K-5 looks really tempting....

it is really tempting, i plan to get it eventually.

for now, i'm happy that the k-r (as an entry level camera) trounces my old k20d in terms of low light performance and af - i can take pictures of cats at night easily for the first time in my life! and that it offers a visible improvement in DR (nearly 1.5 stops more, as tested in DXOmark). pentax has come a long way since i joined it... the k100d's DR was really crippled in some sense.

 

[henry soh]

Hi, Before buying it, I will test the IQ and satisfy myself that it is ideal. One factor alone is not enough to justify purchase. All factors contributing to an ideal IQ must be considered in a system like the camera, including the features that you like to play with. If there is a weak link
in the system, the IQ would not be satisfactorily displayed. The best camera will give you an ideal IQ even in difficult lighting conditions.

 

[LBL2009]

How will having a better dynamic range helps? Thanks.

A better dynamic range allows me to keep details in the highlight without losing much shadow or the other way round. And I will have less overexposed if I shoot in one of the auto mode.

I am very keen to see how K5's 14 EV shows up in a photo against another camera with lower EV. :think:

 

[coolthought]

sunset timing, DR needed is wider?

granted, you might still need to use tools like multiple exposures for blending, and gnd.... but it offers opportunities to just do this with a single shot with gnd alone (more often).

A better dynamic range allows me to keep details in the highlight without losing much shadow or the other way round. And I will have less overexposed if I shoot in one of the auto mode.

I am very keen to see how K5's 14 EV shows up in a photo against another camera with lower EV. :think:

thank you guys for your insight.
Anyone with a K-5 to show us how good it is?

 

[doodah]

Anyone with a K-5 to show us how good it is?

I'll take this one step further: can anyone show side-by-side comparison test images from K-5 and K20D to demonstrate the 'useful' extra dynamic range?

The D90/D5000 are believed to have at least 1 stop "more dynamic range" than the D300s which was a later release. Do you honestly believe that is possible, given ALL are 12 MP APS-C sensors? Also, why will Nikon choose to release a later semi-pro camera (D300s announced on 30Jul09) that has a WORSE sensor than its cheaper sibling (D90 announced on 17Aug08)?... Some food for thought...

Anyway, have a good read here, here and here. This is why I've always insisted in getting two cameras side-by-side and carry out careful, detailed testing. Real world images tell you a lot more than some silly useless tests...

 

[alantkh]

anyone wants to meet up to try?

It should not take more than 10 minutes. Mount your cameras on a tripod. Shoot at manual at various exposure settings at the same time for all cameras. See where your camera clip highlights and when the noise overwhelm the signal?

I can contribute a D700 for the test. I have a 24mm lens. I am particularly interested in how the k-5 stacks up.

 

[LBL2009]

D7000 and K5 use the same Sony sensor? They score high and are close in the DXOmark tests.

New model like the Kr and A55 also tested with high scores.

This seems to suggest an advancement in sensor technology. Good news.

The benchmark is raised. :thumbsup:

 

[night86mare]

I'll take this one step further: can anyone show side-by-side comparison test images from K-5 and K20D to demonstrate the 'useful' extra dynamic range?

i haven't had the chance to try out the k-5, but i can tell you very quickly that 3 stops can make a difference, but the increment makes less and less difference as you move up the scale. for example, 17 stops DR vs 14 stops DR would probably be less useful than an increment of 3 stops from 11 stops DR.

the k100d versus k20d, when i was using the k100d, i used exactly the same set of equipment (outside of camera body + lens), i.e. tianya gnd filters. i always ended up having to go through the painful HDR process (which gave subpar output, imho) with the k100d... which i hardly ever needed with the k20d with careful exposure. with an additional 1.5 stops with the k-r, it is a given that i can throw photomatix into a small closet that would only be unlocked when there is an extremely compelling reason to whip it out again... maybe once a year. :devil:

that said, i agree with you, most people would not see a difference between k20d and k5. but if it's there, it's there, and it can come in useful when you need it.

 

[LBL2009]

This owner of a hardworking 11EV D40 knows too well what difference a one stop (not to mention 2 stops) dynamic range means to him when he shoots landscape in high contrast light situation.

14 EV may not mean much to a 12 EV camera photographer if he has not tried it. If he tries and the 14 EV delivers, I think he will want it. :bsmilie:

 

[doodah]

the k100d versus k20d, when i was using the k100d, i used exactly the same set of equipment (outside of camera body + lens), i.e. tianya gnd filters. i always ended up having to go through the painful HDR process (which gave subpar output, imho) with the k100d... which i hardly ever needed with the k20d with careful exposure.

But were you shooting the exact same scene with the two cameras side-by-side? Also, was it really possible to uncover more highlights and shadow details from one compared to another? If it was a problem with shadow details, could shadow NR be performed on the camera with more noisy output so that it matches the other? Are you seeing results like that?

 

[night86mare]

But were you shooting the exact same scene with the two cameras side-by-side? Also, was it really possible to uncover more highlights and shadow details from one compared to another? If it was a problem with shadow details, could shadow NR be performed on the camera with more noisy output so that it matches the other? Are you seeing results like that?

oh, no.

but believe me, when you've shot enough sunsets you'd know. my entire workflow changed, because there were possibilities.

to your second question, clearly yes - given the same scene i was not able to uncover sufficient detail from a single exposure for sure.

 

[coolthought]

I'll take this one step further: can anyone show side-by-side comparison test images from K-5 and K20D to demonstrate the 'useful' extra dynamic range?

The D90/D5000 are believed to have at least 1 stop "more dynamic range" than the D300s which was a later release. Do you honestly believe that is possible, given ALL are 12 MP APS-C sensors? Also, why will Nikon choose to release a later semi-pro camera (D300s announced on 30Jul09) that has a WORSE sensor than its cheaper sibling (D90 announced on 17Aug08)?... Some food for thought...

Anyway, have a good read here, here and here. This is why I've always insisted in getting two cameras side-by-side and carry out careful, detailed testing. Real world images tell you a lot more than some silly useless tests...

i haven't had the chance to try out the k-5, but i can tell you very quickly that 3 stops can make a difference, but the increment makes less and less difference as you move up the scale. for example, 17 stops DR vs 14 stops DR would probably be less useful than an increment of 3 stops from 11 stops DR.

the k100d versus k20d, when i was using the k100d, i used exactly the same set of equipment (outside of camera body + lens), i.e. tianya gnd filters. i always ended up having to go through the painful HDR process (which gave subpar output, imho) with the k100d... which i hardly ever needed with the k20d with careful exposure. with an additional 1.5 stops with the k-r, it is a given that i can throw photomatix into a small closet that would only be unlocked when there is an extremely compelling reason to whip it out again... maybe once a year. :devil:

that said, i agree with you, most people would not see a difference between k20d and k5. but if it's there, it's there, and it can come in useful when you need it.

I saw the photos taking with the K-5 during a recent pentax outing. Here.
Really impressive. Gives a even better and natural feel than HDR.